Pragmatic Designs for Combining Cyclic Laboratory and Monitor Data.

Area of Science:

• Biostatistics
• Biomedical Engineering
• Endocrinology

Background:

• Cyclic biologic processes, like hormone fluctuations, require accurate measurements.
• Laboratory measurements offer high accuracy but are costly and infrequent.
• Home monitor data is less expensive but less precise, often analyzed separately.

Purpose of the Study:

• To propose a cost-effective study design combining inexpensive and laboratory measurements for cyclic biologic processes.
• To enhance the efficiency of effect estimates in studies of cyclic biological data.
• To compare the proposed design's efficiency and cost benefits against alternative measurement strategies.

Main Methods:

• A novel design integrating frequent inexpensive monitor data with sparse, strategically sampled laboratory data.
• Laboratory sampling times determined using a D-optimality criterion for maximum information gain.
• Comparison of the proposed design against designs using only laboratory measurements or equally spaced laboratory measurements.

Main Results:

• The proposed combined design achieves efficiency comparable to using all laboratory measurements.
• Significant performance improvement over designs with equally spaced laboratory measurements.
• Model robustness demonstrated under limits of detection (LOD), with largely unbiased estimates.

Conclusions:

• Combining inexpensive and laboratory measurements offers a cost-effective and efficient approach for studying cyclic biologic processes.
• The D-optimality criterion effectively optimizes laboratory sampling for enhanced data analysis.
• This integrated design provides substantial benefits for analyzing hormone trajectory data, such as in menstrual cycles.